60.nfs

[metze/ctdb/wip.git] / config / events.d / 60.nfs

diff --git a/config/events.d/60.nfs b/config/events.d/60.nfs
index f424f8cb2951759270a3bd59c98af2166e5d8ae5..87955df92ccd4595210943f734d83a1f0afe657c 100755 (executable)
--- a/config/events.d/60.nfs
+++ b/config/events.d/60.nfs
@@ -17,8 +17,12 @@ service_stop="startstop_nfs stop"
 
 loadconfig
 
+[ "$NFS_SERVER_MODE" != "GANESHA" ] || exit 0
+
 ctdb_start_stop_service
 
+is_ctdb_managed_service || exit 0
+
 case "$1" in 
      init)
        # read statd from persistent database
@@ -78,15 +82,20 @@ case "$1" in
                (
                        service_name="nfs_knfsd"
                        ctdb_counter_incr
-                       ctdb_check_counter_limit 10 quiet >/dev/null
+
+                       ctdb_check_counter_equal 10 || {
+                               echo "Trying to restart NFS service"
+                               startstop_nfs restart >/dev/null 2>&1 &
+                               exit 0
+                       }
+
+                       ctdb_check_counter_limit 15 quiet >/dev/null
                ) || {
                        echo "$ctdb_check_rpc_out"
                        echo "Trying to restart NFS service"
                        startstop_nfs restart
                        exit 1
                }
-               # we haven't hit the failure limit so restart quietly
-               startstop_nfs restart >/dev/null 2>&1 &
            fi
        }
 
@@ -98,16 +107,51 @@ case "$1" in
        } || exit $?
 
        # check that lockd responds to rpc requests
-       ctdb_check_rpc "LOCKD" 100021 1 || {
-               echo "Trying to restart lock manager service"
-               startstop_nfs restart
-               startstop_nfslock restart
-               exit 1
-       }
+       if ctdb_check_rpc "LOCKD" 100021 1 >/dev/null ; then
+               (service_name="lockd"; ctdb_counter_init)
+       else
+               (
+                       service_name="lockd"
+                       ctdb_counter_incr
+
+                       ctdb_check_counter_equal 10 || {
+                               echo "Trying to restart NFS lock service"
+                               startstop_nfs restart >/dev/null 2>&1 &
+                               startstop_nfslock restart  >/dev/null 2>&1 &
+                               exit 0
+                       }
+
+                       ctdb_check_counter_limit 15 quiet >/dev/null
+       ) || {
+                       echo "$ctdb_check_rpc_out"
+                       echo "Trying to restart NFS lock service"
+                       startstop_nfs restart
+                       startstop_nfslock restart
+                       exit 1
+               }
+       fi
 
        # mount needs special handling since it is sometimes not started
        # correctly on RHEL5
-       ctdb_check_rpc "MOUNTD" 100005 1 || {
+       if ctdb_check_rpc "MOUNTD" 100005 1 >/dev/null ; then
+               (service_name="nfs_mountd"; ctdb_counter_init)
+       else
+       (
+               service_name="nfs_mountd"
+               ctdb_counter_incr
+
+               ctdb_check_counter_equal 5 || {
+                       p="rpc.mountd"
+                       cmd="${p}${MOUNTD_PORT:+ -p }${MOUNTD_PORT}"
+                       echo "Trying to restart MOUNTD [${cmd}]"
+                       killall -q -9 $p
+                       $cmd &
+                       exit 0
+               }
+
+               ctdb_check_counter_limit 10 quiet >/dev/null
+       ) || {
+               echo "$ctdb_check_rpc_out"
                p="rpc.mountd"
                cmd="${p}${MOUNTD_PORT:+ -p }${MOUNTD_PORT}"
                echo "Trying to restart MOUNTD [${cmd}]"
@@ -115,6 +159,9 @@ case "$1" in
                $cmd &
                exit 1
        }
+       fi
+
+
        # rquotad needs special handling since it is sometimes not started
        # correctly on RHEL5
        # this is not a critical service so we dont flag the node as unhealthy
@@ -128,10 +175,9 @@ case "$1" in
 
        # once every 60 seconds, update the statd state database for which
        # clients need notifications
-       LAST_UPDATE=`stat --printf="%Y" $CTDB_VARDIR/state/statd/update-trigger`
+       LAST_UPDATE=`stat --printf="%Y" $CTDB_VARDIR/state/statd/update-trigger 2>/dev/null`
        CURRENT_TIME=`date +"%s"`
-       expr "$CURRENT_TIME" ">" "(" "$LAST_UPDATE" "+" "60" ")" >/dev/null 2>/dev/null
-       [ $? = "0" ] && {
+       [ $CURRENT_TIME -ge $(($LAST_UPDATE + 60)) ] && {
            mkdir -p $CTDB_VARDIR/state/statd
            touch $CTDB_VARDIR/state/statd/update-trigger
            $CTDB_BASE/statd-callout updatelocal &